Rebirth 08: Rise from copycat phones

Chapter 537 The Embarrassment of Four Star Semiconductors
The listing of Zhuyun Group in China has had a significant impact, attracting a large influx of deposit funds into the stock market, which will have a far-reaching macroeconomic impact!

Even disregarding external changes, the impact of the $300 billion raised into Zhiyun Group was immense.

With a large influx of funds into Zhiyun Group, the company is not going to hold back its spending... As a top-tier enterprise of such a massive scale, Zhiyun Group is very good at spending money. This money will be used for the development and upgrading of the group's existing and future businesses, as well as a large amount of research and development.

This will directly impact upstream and downstream industrial chains, thereby creating a large number of middle- and high-income jobs.

Just after Zhuyun Group completed its domestic listing, its subsidiary Zhuyun Microelectronics announced that it will invest in building an advanced semiconductor factory in Tongcheng: Zhuyun Microelectronics Tongcheng Production Base Plant No. 38.

"This is a top-tier semiconductor factory designed for the future, used to manufacture equivalent 3-nanometer processes. The main building has been completed, and the next step will be equipment installation and commissioning. We hope to complete the initial construction plan and conduct technology verification next year, and strive for mass production the year after next!"

This information was revealed to the outside world by Ding Chengjun, CEO of Zhiyun Microelectronics. Although the information disclosed was limited, such as the investment amount and expected production capacity, it was not mentioned.

However, this is still the first time that Zhuyun Group has officially announced its construction plan for an equivalent 3-nanometer process factory to the outside world.

Prior to this, Zhiyun Group had only talked about future 3-nanometer processes and technological advancements, but had not announced detailed factory construction plans or expected chip application practices. Ding Chengjun, however, has given the outside world a clear signal this time.

Our 3-nanometer semiconductor factory has already started construction and is expected to be put into mass production within two years.

This news has also put immense pressure on Zhiyun Microelectronics' main competitors, TSMC, Quadstar Semiconductor, and Intel.

Zhiyun Microelectronics' process technology is advancing too fast, and they have too much money. They can develop the three-nanometer process without hesitation!

This is no small matter. According to current estimates in the semiconductor industry, the construction cost for a production capacity of 10,000 wafers per month is expected to reach more than four billion US dollars. This is still talking about domestic production. If it were to be built overseas, such as in the United States, the cost would likely increase significantly due to various factors, reaching five to six billion US dollars per 10,000 wafers.

Based on domestic costs, if Zhiyun Microelectronics were to build a factory with a monthly capacity of 50,000 wafers, the investment cost for just the construction of a semiconductor factory would be over 20 billion US dollars.

This doesn't even include research and development costs.

Such enormous investment costs have already made many semiconductor manufacturers feel overwhelmed in the field of advanced semiconductor manufacturing.

However, even if we are powerless, we have to grit our teeth and keep going. Otherwise, if we ignore everything, Zhiyun Microelectronics will dominate the market and completely monopolize the high-end semiconductor manufacturing business in the next few years.

They can't very well let Zhiyun Microelectronics manufacture high-performance chips for Apple and Qualcomm... Even if Zhiyun Microelectronics were willing, they wouldn't feel comfortable with it.

Therefore, TSMC quickly followed suit with its 3nm process plan, while Samsung was in a somewhat awkward position, as its 7nm process had only been developed for a short time and its yield and production capacity were still not very good.

As for the 5nm process, Zhiyun Microelectronics has already completed technology verification, and even TSMC has announced that it has completed the technology verification of the equivalent 5nm process, but Sixing Semiconductor's 5nm process is still nowhere to be seen...

There are reasons why Samsung Semiconductor is lagging behind, the main reason being that they lack foundry orders for advanced semiconductors, and they do not have related advanced chip businesses themselves.

Taking the 7nm process market as an example, there are currently three main product types: mobile phone SoCs, AI chips, and computer CPUs.

The main manufacturers of mobile phone SoCs are Zhiyun Semiconductor, Apple, and Qualcomm, plus Huawei, which is a relatively small player.

The orders from Zhiyun Semiconductor and Huawi were all placed with Zhiyun Microelectronics.

Fruit orders, as well as some Qualcomm orders, are delivered to TSMC.

After Samsung developed its 7nm process, it only secured a small number of Qualcomm manufacturing orders...

Then there are PC manufacturers (CPs). There are three players in this business: Wistron Corporation (a subsidiary of the Zhuyun Group), Intel, and AMD. Wistron Corporation's orders go to Zhuyun Microelectronics. Intel has its own factory, while AMD's orders are mostly given to TSMC, with a small amount going to Zhuyun Microelectronics and a small amount to Fourstar Semiconductors.

There's also a consumer-grade GPU, which is a gaming graphics card. There are only two players: Zhiyun Semiconductor and AMD. Zhiyun Semiconductor's orders are placed with Zhiyun Microelectronics, while AMD's orders are placed with TSMC. Four Star Semiconductors doesn't have many orders.

Finally, there is another major consumer of production capacity, and that is the computing power chip sector!
This includes server GPUs, namely the AI/APO series graphics cards, as well as EYQ terminal chips, PX terminal chips, ZY big data processing chips, etc. In the field of computing power chips, Zhiyun Semiconductor occupies 99% of the market share, and all of Zhiyun Semiconductor's orders are produced by Zhiyun Microelectronics.

In the field of computing power chips, there are also weak players like AMD and Qualcomm, who also have some terminal computing power chips, but the volume is not large. These small orders are given to TSMC.

Samsung Semiconductor's 7-nanometer process manufacturing has never even produced a computing chip!

The production capacity of the aforementioned advanced chips accounts for more than 95% of the current production capacity demand for the 7-nanometer process (Intel's 10-nanometer process).

Zhiyun Microelectronics manufactures and sells its own products, capturing the vast majority of the market share.

TSMC, as a professional foundry, has captured a small portion of the market share.

Intel manufactures and sells its own products on a small scale, and focuses solely on the CPU field, so there's not much to discuss there.
Four Star Semiconductor does not have its own chip production capacity, and it cannot compete with TSMC in foundry services, so it can only take on some leftover foundry orders.

Then Four Star Semiconductor found itself in a bit of an awkward situation... because it wasn't easy to get orders for manufacturing scrap parts.

The current 7nm process is extremely expensive to fabricate and has high manufacturing costs. At present, there are not many semiconductor design companies that can afford to use the 7nm process. Most chip design companies today have focused their main chip manufacturing processes on the 14nm process node...

After all, some advanced processes at this process node, such as the Zhiyun 10nm and Zhiyun 12nm processes, have relatively good performance and are sufficient to meet the performance requirements of most products. Even the SOCs used in current mid-to-low-end mobile phones are still at this process node.

Moreover, the cost of the 14-nanometer process node has gradually decreased, while production capacity remains sufficient.

Therefore, most semiconductor design companies choose the 14-nanometer process node when designing chips, and chips using the 7-nanometer process node are very rare.

This means that Samsung Semiconductor's 7-nanometer process is struggling to find enough customers... which is quite embarrassing.

There aren't many customers for this 7-nanometer process, so the profits are naturally not high, making it even more difficult to make a large investment to build a factory.

Without money, what's the point of playing with semiconductors!

Therefore, the five-nanometer process at the four-star site can still be pushed forward despite the difficulties, since it can be advanced based on the existing seven-nanometer process technology without incurring too much cost. However, the equivalent three-nanometer process is truly impossible to advance; the technical difficulty is too high, building a factory is too expensive, and even if it were developed, there would be no customers…

Future orders for 3-nanometer process technology will be fewer and more concentrated!

In fact, it's not just Samsung Semiconductor that's struggling to get orders; even TSMC's expected 3nm orders are few and far between... They're only being supported by future advanced mobile SoC orders from Apple and Qualcomm.

However, they don't get many orders for computing power chips, which consume even more production capacity.

Unfortunately, the computing chip industry is dominated by Zhiyun Group. There are no alternative products, and they would rather have a global shortage and supply falling short of demand than find external competitors to manufacture their server GPUs.

The situation across the entire semiconductor industry chain, from design to manufacturing, has made it extremely difficult for Four Star Semiconductors to continue advancing its future 3-nanometer process.

Of course, despite the numerous challenges of future 3nm processes, Four Star Semiconductors has still announced their future 3nm process plans. As for when they will be implemented, we'll see...

Investing in and building a factory with equivalent 3-nanometer technology is just one part of Zhuyun Microelectronics' massive investment plan.

"Our planned production capacity for the future 3-nanometer process is quite large, because our mobile SoC and computing chip businesses both require a large amount of advanced semiconductor production capacity!"

"Especially our computing chips, which will require a huge monthly production capacity of 200,000 or even 300,000 units in the future. After all, the demand for our AI/APO series graphics cards is too large. Not only is the demand huge for enterprises in the field of artificial intelligence, but the computing power demand for our virtual devices is also extremely large. Then there is also the ZY series chip and the EYQ chip, and the production capacity demand for these two types of chips is also not small!"

"In fields such as mobile phone SOCs and computer CPUs, we expect the current situation of 7-nanometer process to continue, gradually transitioning from mature process to advanced process, and generating huge demand for production capacity!"

"We predict that the mobile phone SOC field will gradually transition to the three-nanometer process stage in the next five years. Based on our Zhuyun Semiconductor's mobile phone SOC shipment volume, we will probably need more than 150,000 three-nanometer process capacity in five years!"

"Therefore, in terms of our capacity planning for the 3-nanometer process, we expect to invest in and build a capacity of 150,000 wafers within three years and a capacity of 300,000 wafers within five years."

"To achieve this, we anticipate needing $120 billion in construction costs over five years in the 3-nanometer process area. Considering inflation and rising costs from other sources, this figure could eventually reach over $140 billion!"

Listening to Ding Chengjun's words, Xu Shenxue remained calm, not at all shocked by the investment plan of more than 100 billion US dollars... This is how money is spent in the advanced semiconductor industry, he is used to it.

The investment projects that were once said to be worth hundreds of billions of dollars have now become worth hundreds of billions of dollars.

He said he could still hold on!

Seeing his boss remain calm and unfazed by the over $100 billion investment plan, Ding Chengjun breathed a sigh of relief. He then picked up another report: "This is our capacity expansion plan for 5-nanometer and 7-nanometer processes!"

"Given the complexity and high manufacturing cost of the future 3-nanometer process, only a few types of advanced chips can afford to use it. However, a large number of other chips will still use the 7-nanometer process. In addition, in the next five years or even five years from now, a large number of chips with mature processes will gradually move from the 14-nanometer process node to the 5-nanometer or 7-nanometer process!"

"We predict that in five years, the 7-nanometer process node will replace the 14-nanometer process node and become the mainstream chip process node. Both the internal demand within the group and the demand from external foundries will be very large."

"For example, the APO4500 graphics card used in our entry-level virtual machine devices is still manufactured using a 12-nanometer process, but this obviously cannot last long and is limited by insufficient 3D advanced packaging capacity."

"However, our 3D packaging capacity will increase significantly next year, and even more the year after. At that time, we will be able to meet the production needs of more advanced computing cards!"

"At this point, we will need to have the corresponding 7-nanometer process capacity to maintain the expansion!"

"Therefore, the capacity expansion of our 5nm and 7nm process nodes is still ongoing. In addition to our original plan of 300,000 wafers, we plan to build two new chip factories to bring the total capacity to more than 400,000 wafers!"

"The expansion of this process node is expected to cost around 30 billion US dollars."

"As for the 14nm process node, our original expansion plan is sufficient. Although the capacity is still relatively tight at present, as the advanced capacity gradually increases, some of the demand for the 14nm process will be transferred to the advanced process. In the future, the 14nm process node will be mainly used for the capacity needs of conventional mature chips within the group and external foundry services!"

"In addition, semiconductor manufacturers such as TSMC, Samsung, GlobalFoundries, UMC, and SMIC are now expanding their 14nm process capacity on a large scale. The foundry price for this process node is falling relatively quickly, and the profit margin is declining rapidly!"

"With this in mind, we have halted our large-scale investment plans at the 14-nanometer process node!"

As for the more mature and outdated 28-nanometer process node, Zhiyun Microelectronics stopped investing in it several years ago, and now it's mostly domestic semiconductor manufacturers playing with it!

Although Zhiyun Microelectronics looks down on the 28-nanometer process node, deeming it to have low profit margins and low earnings, the market size is still very large, and it is still quite profitable.

In particular, the domestic semiconductor equipment and consumables are already very mature. Gulf Technology's DUV immersion lithography machine has sufficient production capacity, is not expensive, and has a high cost performance!

Furthermore, Fairy Mountain Holdings also provides wafer fab package services... As long as the customer pays, Fairy Mountain Holdings directly takes care of the entire wafer fab equipment, provides construction technology solutions, teaches you how to build a wafer fab step by step, and also gives away many patent licenses in the semiconductor manufacturing field from Zhiyun Microelectronics.

Patents for processes at 65 nanometers and above are given away for free; use them freely!
Patents for 28-45 nanometer process nodes are also negotiable. As long as you are willing to spend the money, Zhiyun Microelectronics can also grant you patent licenses... For Zhiyun Microelectronics, in mature process fields, sometimes collecting patent fees is more profitable than directly manufacturing chips!
Of course, we won't sell the 14-nanometer process node below 28-nanometer for the time being. The 14-nanometer process node is the main source of revenue for Zhiyun Microelectronics, and it's impossible to sell off the patents now and create more competitors.

However, the 28-nanometer process is sufficient for many second- and third-tier semiconductor manufacturers and can be used in many fields. Its market demand remains strong, with an annual market size of around 20 billion US dollars.

Therefore, in the past two years, many semiconductor manufacturers have emerged in China with 28-nanometer process factories and construction plans, and their production capacity has gradually increased!
China's mature semiconductor industry chain is developing extremely rapidly, and its scale is increasing rapidly... Moreover, prices are starting to fluctuate quite sharply.

This forced Zhiyun Microelectronics to continuously lower prices at the 28-nanometer process node.

Other major semiconductor manufacturers were also forced to lower their prices one after another.

This large-scale price reduction also means that mature semiconductor fields worldwide, such as many sensor, automotive, and home appliance chips, are beginning to gradually migrate from 65-nanometer and 45-nanometer process nodes to 28-nanometer process nodes.

This has also led to an expansion of the market capacity for the entire 28-nanometer process!
Although Zhiyun Microelectronics doesn't value the 28nm process, it still has great potential, and many second- and third-tier semiconductor manufacturers are investing heavily in expanding their production capacity.

A prime example is SMIC, which has made significant investments in the 28-nanometer process node in an attempt to capture a larger market share!

However, these things have little to do with Zhiyun Microelectronics.

Zhiyun Microelectronics is eyeing 7 nanometers, and even 3 nanometers and 2 nanometers in the future. They are playing with the most advanced and cutting-edge processes to earn the most substantial profits!
(End of this chapter)